A naturalistic study of prospective memory function in MCI and dementia

A naturalistic study of prospective memory function in MCI and dementia
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  1 British Journal of Clinical Psychology (2011) C   2011 The British Psychological Society  TheBritishPsychologicalSociety A naturalistic study of prospective memoryfunction in MCI and dementia Claire L. Thompson 1 , 2 , 3 , 4 ∗ , Julie D. Henry 1 ,Adrienne Withall 2 , 3 , 4 , Peter G. Rendell 5 and Henry Brodaty 2 , 3 , 4 1 School of Psychology, University of New South Wales, Sydney, Australia 2 Dementia Collaborative Research Centre, University of New South Wales,Sydney, Australia 3 School of Psychiatry, University of New South Wales, Sydney, Australia 4 Academic Department for Old Age Psychiatry, Prince of Wales Hospital, Sydney,Australia 5 School of Psychology, Australian Catholic University, Melbourne, Australia Objective.  Naturalistic measures of prospective memory (PM) show less age-relateddecline than laboratory measures. We investigated whether a naturalistic measure of PMdifferentiates between normal ageing, mild cognitive impairment (MCI), and dementia. Method.  Ninety-eight older adults agreed to perform a time-based PM task in theireveryday lives. Results.  Despite a self-selection bias in task acceptance, dementia participantsperformed more poorly relative to both the MCI and control group. Performance onthe naturalistic PM task showed good convergent validity with both a cognitive screeningmeasure and a laboratory PM assessment. Conclusions.  PM difficulties are experienced in the everyday lives of people withdementia and are related to laboratory-based assessments but do not appear to beevident on a naturalistic task for those with MCI. Prospective memory (PM) is memory for future intentions and is especially important inlate adulthood, when functional independence is contingent on the ability to remember activities, such as taking medication, turning off appliances, and paying bills. Mildcognitive impairment (MCI) is a syndrome defined as cognitive decline greater thanexpected for an individual’s age without any functional impairment in activities of daily living. The amnestic subtype of MCI has a high risk of progression to Alzheimer’sdisease and MCI is therefore regarded as a risk state for dementia (Gauthier   et al. , 2006;Ritchie, 2004). The diagnostic category of MCI is argued to be clinically useful to identify individuals at an early point in cognitive decline to target preventative and therapeutic ∗ Correspondence should be addressed to Claire L. Thompson, School of Psychology, University of New South Wales, Sydney,2052, Australia (e-mail: DOI:10.1111/j.2044-8260.2010.02004.x  2  Claire L. Thompson et al. strategies as they become available (Petersen, 2004). It is well known that retrospectivememory is impaired in dementia (McKhann  et al. , 1984; Spaan, Raaijmakers, & Jonker,2005) and in MCI (Artero, Petersen, Touchon, & Ritchie, 2006). Laboratory studieshave shown PM to also be impaired in these populations (Duchek, Balota, & Cortese,2006; Jones, Livner, & Backman, 2006; Kinsella, Ong, Storey, Wallace, & Hester,2007).The importance of naturalistic ecologically valid measurement of memory functionhas long been recognized (Bruce, 1985; Koriat & Goldsmith, 1994; Latham, 1978) but isparticularly important in the case of PM (Phillips, Henry, & Martin, 2008). An essentialfeature of PM is the presence of an ongoing task in which the prospective intention isembedded. The nature of the ongoing task has an impact on PM performance (Marsh & Hicks, 1998; Marsh, Hicks, & Cook, 2005) and will vary with the setting of thetask (naturalistic vs. laboratory). A key defining criteria of naturalistic PM tasks is that theongoing task is embedded in the participants’ usual daily activities and is therefore under thecontroloftheparticipant.Thiscontrastswithlaboratory-basedassessment,wheretheexperimenter controls the setting and the ongoing task (Bailey, Henry, Rendell, Phillips,& Kliegel, 2010). Retrospective memory is not embedded in an ongoing activity andconsequently is equally effortful and difficult across laboratory and naturalistic settings(Kliegel, Rendell, & Altgassen, 2008), whereas PM may be more effortful in one contextthan in another because the setting and the ongoing task have an influence on PM task difficulty. Yet to date, only one study naturalistic study of PM in cognitively impairedolder adults has been conducted (Will  et al. , 2009).Early PM studies used measures with high ecological validity such as askingparticipants to return postcards on certain days or telephone the experimenter atcertain times (Harris, 1984). A shift towards controlled laboratory studies then ensued,typically involving computer-based tasks such as pressing a specified key in responseto a cue on screen (e.g., Einstein, Smith, McDaniel, & Shaw, 1997). While undoubtedly important in clarifying key characteristics and correlates of PM function, the ecological validity (and real-life generalizability) of such tasks is questionable (Kliegel, McDaniel, & Einstein, 2000). The contrast between findings from laboratory and naturalistic settings was strikingly illustrated in a meta-analytic review that revealed older adults performsubstantially worse than younger people in laboratory settings but substantially better than younger counterparts in naturalistic studies (Henry, Macleod, Phillips, & Crawford,2004; Phillips  et al. , 2008).Phillips  et al.  (2008) identified two essential criteria for naturalistic PM tasks:  context  (PM task is embedded in daily life of participants) and  time period   (typically conductedover several days). They classified naturalistic studies of PM into five types of ecological validity.Typeone(themostecologicallyvalid)entailsboththetasksettingandtheactionrequiredbeingapartoftheparticipant’susualenvironmentandroutine.Thiscanonlybeachievedthroughobservationofbehaviour,notbyexperimentation.Typetwoecological validity involves embedding an experimental task into the participant’s everyday routineandenvironment.Typesthree,four,andfivearelowerlevelsofecologicalvaliditywherethe task is set in an artificial environment. Studies using Type two ecological validity (to which the present study belongs) have provided evidence of age-related improvementsinPMcomparedtoyoungerparticipants.Suchstudieshavenotyetbeenconductedwith participants diagnosed with MCI or dementia.Naturalistic assessment of PM is essential to ascertain the degree to which any effectsobserved in laboratory settings generalize to everyday life and to assess the validity of existing laboratory assessments (Phillips  et al. , 2008). This is particularly relevant  Naturalistic PM in MCI and dementia  3 for cognitive disorders of old age: specifically, MCI and dementia. This is because PMlapses are considered to be amongst the most distressing and disabling features of both disorders. For instance, it has been found that PM failures cause more difficulties in daily living and caregiver burden than do retrospective memory failures (Smith, Della Sala,Logie, & Maylor, 2000).Only one study to date has examined cognitively impaired older adults usingboth laboratory and naturalistic measures of PM (Will  et al. , 2009). In this study,cognitively impaired older adults performed more poorly than age-matched controlson both measures with the magnitude of the deficit comparable across settings. Thisimplies that PM deficits observed in laboratory settings may be a valid indicator of PM difficulties in everyday life. However, the sample size in this study was relatively small (  n  =  15 per group) and no formal diagnostic criteria were used to classify impairment. Whilst PM difficulties are known to occur in dementia and in MCI, the criteria for MCI specify the absence of functional impairment. Yet PM failures cause functionalimpairment. Based on laboratory findings, PM impairment would be expected on anaturalistic task. However, the criteria of absence of functional impairment suggestnaturalistic PM should be intact in MCI. The first aim of the present study was thereforeto empirically assess whether a naturalistic measure of PM differentiates betweenparticipants who meet formal diagnostic criteria for MCI and dementia, relative toage-matched controls. The current study is a follow-up of a previously reported study (Thompson, Henry, Rendell, Withall, & Brodaty, 2010) that assessed laboratory PMperformance. The current study assessed naturalistic PM and involved 70% of the samplein the previous study. Thus, the second aim was to assess the convergent validity of naturalistic and laboratory-based PM assessments in MCI and dementia: specifically by assessingperformanceonanaturalisticPMmeasureinasampleofparticipantswhohavealready completed a validated laboratory measure of this construct that has documentedsensitivity to MCI and dementia (Virtual Week; Rendell & Craik, 2000; Thompson  et al. ,2010),aswellastheMiniMentalStateExamination(MMSE;Folstein,Folstein,&McHugh,1975), a widely used screening tool for dementia. Method Participants Participants were recruited into a PM study (  n  =  140; Thompson  et al. , 2010) froma large epidemiological study of ageing that commenced 2 years prior to the currentstudy or from a Memory Disorders Clinic. All participants had been previously assessedand diagnosed as having dementia, MCI, or healthy cognition by consensus conferenceof either the epidemiological study or the memory clinic (using identical test batteriesand criteria, overseen by the same consultant psychiatrist and neuropsychologist) inthe 2 months prior to recruitment into the PM study. All participants were community dwelling,hadadequateeyesight,hearingandEnglishlanguageabilitytocompleteassess-ments, and regular contact with an informant. Participants with previous psychiatric or neurologicalillnesswereexcluded.Of 140recruitedparticipants, 98(70%)acceptedthenaturalistic PM task. Of these, 22 met the  Diagnostic and Statistical Manual of Mental  Disorders  – 4th Ed.  (DSM-IV)  criteria for dementia, 31 met the modified criteria for MCI(Artero  et al. , 2006; Petersen, 2007) and 45 were controls without cognitiveimpairment.  4  Claire L. Thompson et al.  Materials and procedure Ethics approval was obtained from South-Eastern Sydney Illawarra Area Health Service –Eastern Section. After providing informed consent, participants had already completedmeasures for a laboratory-based PM study (Thompson  et al. , 2010) including the MMSEand Virtual Week  (Rendell & Craik, 2000), a computerized laboratory assessment of PMthat has been shown to have good reliability and validity (for a review, see Rendell& Henry, 2009). In Virtual Week, participants move through a series of virtual ‘days’ with tasks that they have to remember to do. Responses are scored identically to thenaturalistic task. Virtual Week was administered on a touch-screen HP Tablet Notebook TC4400 (see Thompson  et al. , 2010).Naturalistic assessment of PM was conducted using Palm R   Z22 handheld electronicorganizers with Experience Sampling Program software (ESP; Barrett & Barrett, 2001).The task was designed to be simple enough to be achievable by all participants. Itinvolved turning the Palm device on once per day for 2 days at an agreed upon pre-specified time and tapping a response box that automatically appeared on the screen. A time-stamp function ensured accurate temporal recording of responses. The task metestablished criteria for a naturalistic PM task in that it was embedded into participants’everyday lives and executed over an extended (2-day) time frame (Phillips  et al. , 2008).Participants were given a choice of six times to complete the task. A choice element was considered important to allow participants to fit the task into their daily routine.Importantly, it has previously been found that time schedule does not affect themagnitude of age effects on a similar time logging task  (Rendell & Thomson, 1999)and no differences were found for choice versus allocated response time conditions(Rendell & Thomson, 1999, experiment 2). Participants were given a demonstration of howtooperatethedevicebyturningitonandusingapenstylustotaparesponsebuttonto answer two multiple choice questions relating to the time of day. Participants thenpracticed the task in front of the examiner, with the demonstration and practice beingrepeated as many times as the participant needed to complete the task independently  with confidence. Instructions given emphasized the challenge of the task was doing itat the agreed upon time of day (as the data to be analysed were extracted from thetime-stamp function of the device). Participants were instructed not to use any writtenor other reminders to do the task. Postage-paid envelopes were provided for return of the device.Responses were scored as follows:  Correct   indicated the response was rememberedat the correct time (defined as within 5 min before or after the agreed-upon time) andscored three points;  Little Late responses were remembered more than 5 min but within20 min after the correct time and scored two points;  Late  responses were more than20 min late and scored one point.  Little Early  (two points) and  Early  (one point) items were the converse of late items.  Missed   indicated the participant did not remember torespond at any time and scored zero points. Each participant received a score for each of the 2 days. Results TherewasagroupbiasintaskacceptancewithcontrolsmorelikelythanMCIordementiaparticipants to accept the task (   p = .018 and .002, respectively) and those with MCI anddementiaequallylikelytoaccept(   p = .437).Thosewhorefusedhadlowerperformancesthan those who accepted on Virtual Week,  t  (138) = 3.71,  p  <  .001, and on the MMSE,  Naturalistic PM in MCI and dementia  5 t  (138) = 2.64,  p = .009.Reasonforrefusaldidnotinteractwithgroupwithmostrefusalsindicating that they   did not want to do the task  (59% of refusals), were  too busy  (12%),or gave  other reasons  (20%). Technical problems with the device accounted for 10% of non-participation. Allocationtothedementia,MCI,orcontrolgroupwasbasedentirelyonthediagnosesmade prior to recruitment into the study. The groups did not differ demographically;age, dementia  M (SD)  =  79.1 (5.68), MCI  =  78.8 (5.48), controls  =  77.7 (4.86);  F  (2,95) = 0.70,  p = .505,  2p = .01; years education, dementia  M (SD) = 11.14 (3.71), MCI = 11.19 (3.48), controls  =  11.53 (3.25);  F  (2, 95)  =  0.18,  p  =  .835,   2p  <  .01; and sex,dementia  =  50.0% male, MCI  =  58.1%, controls  =  37.8%,   2 (2,  N   =  98)  =  3.14,  p  = .208;   = .12). MMSE scores differentiated the groups, dementia  M (SD) = 26.8 (2.53),MCI = 27.7 (1.50), controls = 28.7 (1.34);  F  (2, 95) = 8.82,  p  < .001,   2p  = .16.  Post-hoc  Tukey tests indicated that the dementia and MCI groups’ MMSE differed from controls(   p s  <  .001 and .026, respectively) but that the MCI and dementia groups did not differ (   p = .283). Virtual Week differentiated the groups,  F  (2, 95) = 10.81,  p  <  .001,  2p = .19, with the dementia and MCI groups’ performance lower than controls (   p s  <  .001 and.006, respectively) but with no significant difference between the MCI and dementiagroups (   p = .356). A 3  ×  2 mixed Analysis of Variance (ANOVA) (between-subjects  group  and within-subjects  time  day 1, day 2) identified main effects of both group,  F  (2, 92)  = 5.15,  p  =  .008,   2p  =  .10, and time,  F  (1, 92)  =  5.03,  p  =  .027,   2p  =  .05, with this latter effect indicating better performance on the first day than the secondday (see Figure 1). The interaction between group and time was not significant(   p  =  .58).  Post-hoc   Tukey tests of group showed the dementia group was impairedrelative to both controls (   p  =  .002; Cohen’s  d   =  0.81), and those with MCI (   p  = .025;  d   =  0.66). MCI and control groups did not differ (   p  =  .377,  d   =  0.21). Figure 1.  Mean score on the naturalistic prospective memory task for controls and participants withmild cognitive impairment (MCI) and dementia (maximum score  =  3). Bars represent one standarderror of the mean.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks